Google Ventures is betting big on the future of developer infrastructure, leading a rapid-fire $10 million Series A for Blacksmith just four months after backing the startup's seed round. The Y Combinator alum promises to slash software build costs by 75% while doubling processing speeds, capturing hundreds of enterprise customers as AI coding agents reshape development workflows.
Blacksmith just closed one of the fastest Series A rounds in recent memory. The developer tools startup secured $10 million from Google Ventures in just 14 days, with Alphabet's investment arm doubling down after leading the company's $3.5 million seed round back in May.
The timing isn't coincidental. As AI coding agents flood the software development landscape, the pressure to ship code faster has never been more intense. Blacksmith has positioned itself right in the middle of this transformation, offering a continuous integration and delivery service that promises to cut compute costs by up to 75% while doubling processing speeds.
"Because we're going the bare-metal route, we have much better control over our economics compared to the hyperscalers," CEO Aditya Jayaprakash told TechCrunch in an exclusive interview. The approach gives the company "abundant control" over margins as it scales.
The numbers back up Google Ventures' confidence. Blacksmith went from $1 million in annual recurring revenue in February to $3.5 million ARR today, powered by over 700 customers including enterprise names like Supabase, Clerk, and Chroma. The San Francisco startup accomplished this growth with just eight employees - a team efficiency that caught GV's attention.
Unlike traditional CI/CD providers that rent generic cloud servers from AWS or Google Cloud, Blacksmith runs on high-performance, gaming-grade CPUs housed in their own facilities. This bare-metal approach lets development teams switch with a single line of code change and start shipping faster within minutes.
The startup emerged from the experiences of its University of Waterloo-educated founders, who previously built large-scale distributed systems at Faire and Cockroach Labs. They witnessed firsthand how costly and unpredictable the build and testing stages of software releases can become. "You would have to spin up hundreds of machines and burn through hundreds of hours of computing power just to test new code before shipping it," Jayaprakash explained.
